22. Introduction to Agentic AI

Agentic AI tools use a large language model to plan and carry out multi-step tasks, such as writing and running code, rather than just answer a question. This section covers the current landscape of coding agents, research and science agents, and the frameworks and protocols behind them; how to set up and responsibly run these tools on the Kempner AI cluster, whether in a terminal or through VS Code; how to serve open-weight models on the cluster’s own GPUs so your prompts and code never leave it; and how to fit agents into a research workflow and keep their output trustworthy. For the data-handling rules that govern all of this, see Security and Compliance.